The Science of the Total Environment 319(2004) 13–25

0048-9697/04/$ - see front matter� 2003 Elsevier B.V. All rights reserved.doi:10.1016/S0048-9697(03)00435-2

The potential ofLolium perenne for revegetation of contaminatedsoil from a metallurgical site

M. Arienzo*, P. Adamo, V. Cozzolino

Dipartimento di Scienze del Suolo, della Pianta e dell’Ambiente, Universita di Napoli Federico II, Via Universita, 100,` `

Portici NA 80055, Italy

Received 24 March 2003; received in revised form 7 July 2003; accepted 9 July 2003

Abstract

A greenhouse study was carried out to determine the possibility of usingLolium perenne for revegetation of soilfrom a former ferrous metallurgical plant(Naples, South Italy) contaminated by Cu, Pb and Zn at levels abovecurrent Italian regulatory limits. Surface soil samples(0–40 cm) from the facility area where raw minerals weredisposed(RM1 and RM2), from a nearby unpolluted cultivated soil(C) as control and a 1:3 mixture of the controlwith the polluted ones(RM1qC and RM2qC) were utilized for the experiment. Revegetation trials were conductedin the greenhouse. At 90 days from seeding, shoot length, chlorophyll content, biomass yield, plant metal uptake andchanges of organic carbon content and metal distribution among soil extractable phases defined by sequentialextraction were determined. In the mixed substrates(RM1qC and RM2qC) concentrations of Cu, Pb and Zn werestill two to three times higher than the Italian regulatory limits. Plants were healthy with 100% survival in allsubstrates, with no macroscopic symptoms of metal toxicity. The high pH of the soil could be one of the mostimportant parameters responsible for the limited plant availability of the metals. On RM1, RM2 and mixed media,plants experienced retarded growth, reduced shoot length and biomass yield and higher total chlorophyll contentcompared to those cropped on the control soil, without any evident phytotoxic symptoms. In RM1 and RM2, theplant contents of Cu(19.3 and 12.6 mg kg ), Pb (0.98 and 0.67 mg kg ) and Zn (99 and 88 mg kg ) werey1 y1 y1

higher than that of plants grown on non-contaminated soil(Cu 10.1, Pb-0.2, Zn 79 mg kg ), but still in the rangey1

of physiologically acceptable levels. The distribution of metals in soil was slightly affected byLolium growth withchanges only regarding the organic-bound Cu and Zn pool, with reduction up to 24%. Results indicated that anacceptable healthy vegetative cover can be achieved on the contaminated soil by the proposed revegetation approachand that metals will remain stable over the study period with slight variation of the more available metal forms.� 2003 Elsevier B.V. All rights reserved.

Keywords: Lolium perenne; Heavy metals; Contaminated soil; Ferrous metallurgical plant; Revegetation; Phytostabilization

*Corresponding author. Tel.:q39-081-253-9173; fax:q39-081253-9186.E-mail address: michele.arienzo@unina.it(M. Arienzo).

14 M. Arienzo et al. / The Science of the Total Environment 319 (2004) 13–25

1. Introduction

For the restoration of heavy metal contaminatedsoils different strategies can be adopted. The mostcomprehensive solution involves excavating thepolluted substratum and replacing it with unpol-luted soil. On a large scale, this option is notfeasible due to the costs involved. A more inno-vative approach could be the in-situ or ex-situpurification of the substratum by means of chem-ical or physico–chemical techniques, which‘extract’ or ‘remove’ the metals from the soil.Disadvantages associated with such technologiesare the increased mobility of the remaining metalsand the changes of the physico–chemical charac-teristics of the treated substratum. A third solutionis the in-situ immobilization of the contaminantsby strong metal-immobilizing agents and subse-quent revegetation of the area. In situ immobili-zation can be achieved by use of lime or of zeolites(Gworek, 1992) and beringite(Vangronsveld etal., 1995), i.e. specific aluminosilicates capable offixing the metals to the soil complex(De Boodt,1991).Green plants have been employed in recent years

as a means of stabilising andyor removing metalsfrom contaminated soils. The revegetation activityaims to stabilize the site, to establish a cover cropthat will prevent dispersion of metal-contaminatedparticles by water or wind erosion and to reducemetal mobility by rhizosphere-induced adsorptionand precipitation processes(Vangronsveld et al.,1991, 1993). The long-term goal is the creation ofa self-sustaining ecosystem that can support pro-ductive land use activities and is aestheticallypleasing. Vegetative stabilization also improves thechemical and biological characteristics of the con-taminated soil by increasing the organic mattercontent, nutrient levels, cation exchange capacityand biological activity. These chemical, biologicaland physical improvements accelerate the devel-opment of a viable nutrient cycle and self-sustain-ing vegetative cover and restore the affected areato some acceptable steady-state condition for sec-ondary land use(Norland and Veith, 1995).For revegetation to succeed, the degree of plant

tolerance to contaminants must be known andselection of appropriate species is critical. The

tolerance level of a particular plant species showsto a metal is a direct function of the chemicalspeciation of the metal in the environment and itsbioavailability along with plant physiological fac-tors. Slight changes in availability can cause theseelements to become toxic. Literature(Pichtel andSalt, 1998) reports thatLolium perenne is a suita-ble species for revegetation of metalliferous waste,producing high dry matter yields and accumulatingelevated amounts of metals from reservoirs in thereadily extractable and soluble forms, fitting theSmith and Bradshaw(1979) definition of a facul-tative metallophyte. In other words, the plantshows a tolerance to metals whose concentrationswere not elevated in the parent soil from whichthe particular variety originated and accumulatesmoderate to high levels of metals in its biomass.A plant’s suitability for a particular site is

measured by evaluating uptake of metals, changesin yield, and changes of basic metabolic processessuch as respiration, photosynthesis, CO -fixation2

and gas exchange(Vangronsveld and Clijsters,1994). Metabolic biomarkers are often early indi-cators of stress and provide information about thedegree of chemical exposure at the organism level(Vangronsveld et al., 1996).Root-induced changes in rhizosphere soil prop-

erties may have significant influence on the mobil-ity and bioavailability of nutrients and traceelements(Marschner, 1986). In the context ofphytoremediation little research has been carriedout to examine plant uptake in comparison withsoil metal chemical pools(Hammer and Keller,2002). Thlaspi caerulescens has been shown to bemore effective in mobilising Zn from less solublethan from the mobile fractions in soils(McGrathet al., 1997), indicating the potential of plant tomodify the distribution of metal contaminantsamong soil geochemical phases. Phytoremediationmust take into account this potential of risks orbenefits of using plants in order to improve thelong-term efficiency of the technique.The former ILVA of Bagnoli Steel Plant(IBSP)

located in the West area of Naples is a dismantledferrous metallurgical site, which had been in oper-ation since 1910 and covers approximately 300 hain the western part of the Fuorigrotta–Bagnoliplain. After reaching a peak of steel production of

15M. Arienzo et al. / The Science of the Total Environment 319 (2004) 13–25

2=10 millions of tons at the end of 1960, the6

facility was closed in 1991 and demolished in1998. In a previous study Adamo et al.(2002a)determined the total content, speciation and mobil-ity of heavy metals in soil of the IBSP’s rawmaterials disposal area. Cu, Co, Cr, Ni, Pb and Znwere above the Italian regulatory levels for soilsof public, private and residential areas(ItalianMinistry of Environment, 1999). At this site, dueto the past steel processing activities and plantdismantling, vegetation is limited to few higherherbaceous plants typical of disturbed areas(Buon-donno et al., 1998). The soils measured low innitrogen, phosphorus and macronutrients(Adamoet al., 2002a). The risk of spreading of metals tothe surroundings by wind erosion and water run-off and the IBSP’s location along a popular andattractive portion of Naples’s sea coast argues foran immediate restoration of the site. Moreover,according to Italian law, the site must be remedi-ated. In the case of the IBSP site the substantialpresence of elements in the mineralogical structureof oxides and silicates(Adamo et al., 2002a)obviates the need for using metal immobilizingagents, making direct revegetation of the site arealistic and feasible alternative to other restorationsolutions.The primary purpose of this study was to eval-

uate, at the greenhouse level, the feasibility of theuse of L. perenne L. for the revegetation of theIBSP’s degraded soil. More specifically the objec-tives were:(i) to evaluate the survival and growthof L. perenne on contaminated soil from the IBSPsite, (ii) to monitor phytoaccumulation of Cr, Cu,Ni, Pb and Zn, and(iii ) to investigate the relativeimportance of plant absorption in influencing thechanges of heavy metal distribution among soilgeochemical phases defined by sequentialextractions.

2. Materials and methods

Two soil samples(RM1 and RM2) were col-lected by mixing the upper 40 cm from profilesexcavated in the zone of the IBSP site designatedfor stoking raw iron ore and minerals. The tailingsare typically heterogeneous, contain large stones,display extreme compaction, low hydraulic con-

ductivity, very low porosity and a lack of structure.All these factors prevent or greatly restrict thedevelopment of a vegetation cover. The RM1 andRM2 soils were air dried, crushed and ground topass a 0.95 mm mesh sieve, mixed at a ratio of3:1 with a cultivated unpolluted soil(termed C),collected in the surroundings of the IBSP site, thusobtaining the RM1qC and RM2qC soils. Asubsample of each soil was sieved with a 2 mmmesh sieve and analysed for: particle-size distri-bution at pH 10 after ultrasonic treatment at 20kHz and 75 W for 15 min; pH in a 1:2.5(wyv)soilydeionised water slurry; organic carbon(Walk-ley and Black method, Nelson and Sommers,1982); carbonates(pressure Dietrich–Fruehlingcalcimeter method, Loeppert et al., 1984); cationexchange capacity(CEC) (BaCl method, Hen-2

dershot and Duquette, 1986); exchangeable bases(atomic absorption spectrometry), total-N (stan-dard Kjeldahl method); available soil phosphorus(Olsen method, Olsen and Sommers, 1982).Selected characteristics of the soils are presentedin Table 1. Total concentrations of Cr, Cu, Ni, Pband Zn were determined by digesting 250 mg ofeach soil(-2 mm) in 1:5 HF:HNO in a Mile-3

stone 900 microwave oven at 600 W for 24 min.Metal concentrations in the extract were deter-mined with a Perkin-Elmer Analyst 700 atomicabsorption spectrometer(AAS) equipped withdeuterium-arc background correction in acidicsamples(pH below 2.0). The accuracy of theanalytical determination was obtained using a BCRstandard reference material(CRM 141R-calcare-ous loam soil).The particulate-bound metal species of the soils

were obtained following the selective sequentialextraction scheme shown in Table 2. The sequen-tial extraction was carried out, in duplicate, on 1g soil (-2 mm) in 50 cm polypropylene centri-3

fuge tubes to facilitate washing of the soil aftereach extraction, thus minimizing any loss of solids.After each successive extraction, the supernatantwas removed by centrifugation for 10 min at 5000rev.ymin. The residue was washed with 10 ml ofdeionised, distilled H O by centrifugation for 102

min at 12000 rpm and the supernatants werecombined, mixed and made up to a known volume.Data analysis was carried out on the average values

16M

.A

rienzoet

al./

The

Scienceof

theTotal

Environm

ent319

(2004)13–25

Table 1Selected physical and chemical properties of the raw(RM1, RM2), control (C) and mixed(RM1qC, RM2qC) soils

Soil Gravel Sand Silt Clay pH OC CaCO3 CEC Mg K Na Total N Available P)2 mm 2–0.02 0.02–0.002-0.002 g kgy1 g kgy1 cmol kgy1

c cmol kgy1c cmol kgy1

c cmol kgy1c g kgy1 g kgy1

% g kgy1 g kgy1 g kgy1

RM1 33 880 99 20 8.4 15 51 9.8 1.2 2.4 3.0 0.21 4.1RM2 37 846 140 14 8.4 8 51 9.8 1.2 3.0 3.8 0.10 4.5C 16 772 161 68 8.2 23 45 21.0 1.6 12.9 8.8 0.25 94RM1qC (3:1 wyw) 30 863 104 31 8.2 16 47 12.4 1.1 6.1 4.6 0.22 47RM2qC (3:1 wyw) 30 814 133 52 8.1 11 49 12.3 1.3 6.8 5.7 0.13 45

17M. Arienzo et al. / The Science of the Total Environment 319 (2004) 13–25

Table 2Sequential extraction scheme for particulate-bound metal speciation used in the present study

Step Species Reagent* Shaking time andtemperature(8C)

1 Exchangeable 10 ml of 1 mol l NH NO(pH 7)y14 3 4 h at 258C

2 Carbonate-bound 25 ml of 1 mol l CH CO Na(pH 5)y13 2 6 h at 258C

3 Easily reducible 20 ml of 0.1 mol l NH OH.HCl in 0.01 mol l HNOy1 y12 3 30 min at 258C

metal oxide-bound4 H O extractable2 2 5 ml of 30% H O (pH 2), 3 ml of 0.02 mol l HNO ;y1

2 2 3 2 h at 858Corganic-bound then add 3 ml of 30% H O(pH 2);2 2 2 h at 858C

cool, add 10 ml of 2 mol l NH NO in 20% HNOy14 3 3 30 min at 258C

5 Residual Residue of step 4 was digested with HF–HNO3

The volume of the reagent was for 1 g of soil sample(-2 mm).*

of the duplicate samples; the duplicates had anaverage S.D.(s) of "5%. Deionised, distilledH O was used in preparing stock solutions of all2

reagents, which were checked for possible tracemetal contamination. All glassware used was pre-viously soaked overnight in 10%(vyv) conc.HNO (16 mol dm ) and rinsed with deionised,y3

3

distilled H O.2

To evaluate metal leachability the toxicity char-acteristic leaching procedure(TCLP) was per-formed (Testa, 1997). A subset of soil sampleswas oven-dried at 1058C for 6 h, ground in anagar mortar and sieved at 0.95 cm. Five gram ofsoil were extracted with 100 ml of 0.1 M aceticacid buffered at pH 5.0 with 1 N NaOH. Followingextraction samples were filtered through a 0.8mmglass fiber filter and acidified with nitric acid topH-2.0. For quality control and assurance, ablank was included for each batch of threesamples.For the revegetation test, each substrate was

rehydrated up to field capacity with a standardnutrient solution(5.0 g NH NO l ; 2.7 g Cay1

4 3

(HPO ).2H O l ; 5.1 g MgSO .7H O l ; 2.5 gy1 y14 2 4 2

K SO l ) and equilibrated for 2 weeks.y12 4

In a greenhouse setting at ambient temperature(15–28 8C) illuminated with natural light,L.perenne was seeded in polystyrene trays consistingof 90 wells (4 cm diameter, 5 cm depth), filledwith each of the five equilibrated substrates andgrown for approximately 2–4 weeks.Eight pre-germinatedL. perenne seeds at a

similar shoot development stage were transferredinto individual 400 cm pots containing 2 kg of3

each substrate with three replicates for a total of15 pots. Pots were watered regularly by capillarityto a level close to field capacity. After transplantingplants were twice fertilized with a nutrient solution(5.0 g NH NO l ; 5.1 g MgSO .7H O l ; 2.5y1 y1

4 3 4 2

g K SO l ).y12 4

Every 15 days shoot length was measured. Priorto harvest, chlorophyll was measured: two discs(0.8-cm diameter) fixed 1 cm leaf area were2

collected at the mid-length of the leaf, from bothmid-rib sides. Discs were immersed in acetone andground in an agar mortar. After centrifugation at4000 rpm for 30 min, chlorophylla (Chla) andb(Chlb) contents were determined by UV–Vis spec-trophotometry(Perkin Elmer Lambda 3B). Theabsorbance of the solution was measured at 470,647 and 664.5 nm. Formulae and extinction coef-ficients used for determination of chlorophyllouspigments were published by Lichtenthaler(1987).After 90 days of growth plants were harvested

by cutting the shoots approximately 1 cm abovethe soil surface. Roots were carefully removedfrom the soil, gently shaken and washed to elimi-nate any attached soil. The dry weight(DW) ofplant parts(roots and shoots) was determined afteroven drying at 708C. Shoots were ground in astainless steel Wiley mill to pass through a 0.85-mm diameter screen. Subsamples of ground shootmaterial(500 mg) were digested in a 6:1 mixtureof concentrated HNOyHF. The digested samples3

were brought to a constant volume with deionisedH O and the digests were analysed by AA for Cr,2

Cu, Ni, Pb and Zn. Quality control for plantanalysis was performed by analysing blanks and

18 M. Arienzo et al. / The Science of the Total Environment 319 (2004) 13–25

Table 3Analytical results for CRM 141R and CTA-VTL-2 standard reference material

Element CRM 141R CTA-VTL-2

Certified Measured(ns9) Diff. * Certified Measured(ns7) Diff. *

Mean R.S.D.** Mean R.S.D.** Mean R.S.D.** Mean R.S.D.**

Cr 195 3.6 153 14.0 y21 1.87 0.45 1.95 0.32 4Cu 46.4 3.9 51.1 10.9 10 18.20 3.2 18.94 3.9 4Ni 103 2.9 111 9.6 8 1.98 0.37 1.90 0.23 y4Pb 57.2 2.1 68.3 10.7 19 22.10 5.6 22.77 4.3 3Zn 283 1.8 249 11.5 y12 43.30 8.5 71.86 6.6 66

Concentrations are in mg kg unless otherwise stated.y1

Diff.s% Difference between measured and certified means.*

R.S.D.srelative standard deviation.**

certified reference material(tobacco leaves BCRCTA-VTL-2, Community Bureau of Reference,Commission of the European Communities).Soil from each pot, separated from roots, was

dried, sieved to-2 mm and analysed for organiccarbon and metal distribution as beforeLoliumcropping.Statistical testing was performed using a single

factor analysis of variance(ANOVA) followed bymultiple comparison testing using Tukey’s test asdescribed in Zar(1984). A significance level ofPs0.05 was selected for determination of statis-tical significance. The degree of association amongvariables was assessed through simple linearregression analysis.

3. Results and discussion

3.1. Data quality control

The accuracy of the analytical methods to deter-mine the total heavy metal concentration of soiland plant material were assessed comparing thecertified and measured values for CRM 141R andCTA-VTL-2 standard reference materials andexpressing the difference as percent of the certifiedvalue(Table 3). Analytical precision, expressed asrelative standard deviation(R.S.D.), was measuredas the ratio between S.D. and respective meanvalues of the reference materials replicate analysis.Considering the results for CRM 141R referencematerial the accuracy was acceptable for Cu, Niand Zn. The concentrations of Cr and Pb were

probably underestimated and overestimated,respectively, by the method used. Hence, theresults for Cr and Pb should be interpreted withcaution, since the ranges and variability of theseconcentrations may rather reflect analytical errorsrather than real trends. Considering the results forCTA-VTL-2 reference material Cr, Cu, Ni and Pbmeasured values fell withiny4yq4% of thecertified values. Greater attention should be paidto Zn because a difference ofq66% was meas-ured. Analytical precision was-14% for CRM141R and -6.6% for CTA-VTL-2, for allelements.

3.2. Soil properties before planting

Concentrations of total Cr, Cu, Ni, Pb and Znin the RM1, RM2, C, RM1qC and RM2qC soilsutilized for the revegetation experiment are givenin Table 4. Soils RM1 and RM2 contained Pb at366–301 mg kg and Zn at 600–679 mg kg ,y1 y1

exceeding the regulatory limits(Pb: 100; Zn: 150mg kg ) established for soils of public, residentialy1

and private areas by the Italian Ministry of Envi-ronment (1999). RM1 also contained concentra-tions of total Cu about threefold the limit, 377 vz.120 mg kg . The C soil contained very lowy1

concentrations of total metals and in the case ofPb and Zn these measured approximately three tofour times lower than those of RM1 and RM2.Mixing of C with RM1 and RM2 in the ratio of1:3 (wyw) reduced in RM1 the concentrations ofCu by 7%, of Pb by 31% and of Zn by 19%,

19M. Arienzo et al. / The Science of the Total Environment 319 (2004) 13–25

Table 4Total heavy metal content of raw(RM1, RM2), control (C), and mixed(RM1qC, RM2qC) soils (-2 mm)

Soil Sample Cr Cu Ni Pb Zn(mg kg )y1 (mg kg )y1 (mg kg )y1 (mg kg )y1 (mg kg )y1

RM1 73 377 81 366 679RM2 97 100 125 301 600C 35 70 47 110 157RM1qC (3:1 wyw) 59 (63 )* 350 (300) 66 (72) 252 (302) 548 (548)RM2qC (3:1 wyw) 70 (81) 80 (92) 80 (105) 207 (253) 490 (489)

Values calculated considering total metal content of each soil and the mixing ratio.*

Table 5TCLP -extractable metals in raw(RM1 and RM2) and control(C) soils (-0.95 cm)a

Soil Sample Cr Cu Ni Pb Zn(mg l )y1 (mg l )y1 (mg l )y1 (mg l )y1 (mg l )y1

RM1 2.1 9.9 3.9 6.8 71.3(0.35) (1.15) (1.00) (1.00) (10.02)

RM2 4.3 74.1 6.8 10.9 69.8(1.25) (12.53) (2.00) (2.06) (12.56)

C 1.5 20.6 1.8 6.9 76.1(1.00) (2.50) (0.75) (1.30) (12.70)

Standard deviation in parentheses.TCLPstoxicity characteristic leaching procedure(Testa, 1997).a

whereas in RM2 the contents of Pb and Zndecreased by 31% and 18%, respectively. Thevalues measured after mixing were close to thoseexpected on the basis of the calculationwC totalcontentq(3=RM total content)y4x. Mixing ofsoils improved other soil chemical properties(Table 1). The mixed media showed increasedcontents of clay and organic carbon and hence theCEC. There was also a very significant(Ps0.05)increase of the exchangeableK concentration ofapproximately three times and of available phos-phorus of approximately 10 times in the mixedmedia. The combination of these improved prop-erties make RM1qC and RM2qC a more favour-able plant growth medium compared to RM1 andRM2. The potential mobility of metals in the RM1,RM2 and C soils was extremely low and metalswere practically immobile(Table 5), falling wellbelow the regulatory limits for the TCLP(Crs5000, Pbs5000 mg l ) (USEPA, 1990) and,y1

with the exception of Pb in RM2 soil, those evenmore restrictive defined for CO enriched water2

leachates test(Crs50, Cus1000, Nis20, Pbs

10, Zns3000mg l ) (Italian Ministry of Envi-y1

ronment, 1999).

3.3. Plant growth

L. perenne grew healthy on all soils withoutshowing any visible symptoms of metal toxicityor nutrient deficiency. Plants grown in the C soilexhibited a better growth and at the end of thegrowing cycle double the biomass and significantlytaller shoots(approx. 30%) were yielded byLol-ium compared to the contaminated substrates(Figs.1 and 2). Although growth was retarded in RM1and RM2 in comparison to RM1qC and RM2qC, no significant differences were observedbetween the raw and the mixed media in terms ofbiomass production. Plants grown on RM2qCdisplayed significantly higher shoot length com-pared with RM2(approx. 15%). The chlorophyllcontent of leaves(Fig. 3) was higher in plantsgrown on contaminated media, with the highestlevels measured in the RM2qC, 2.66 mg g .y1

This observation has been observed by others in

20 M. Arienzo et al. / The Science of the Total Environment 319 (2004) 13–25

Fig. 1. Total dry biomass of shoots and roots per pot ofL.perenne grown in raw (RM1, RM2), control (C) and mixed(RM1qC, RM2qC) soils. Values within the same columnfollowed by the same letter are not statistically significant atPs0.05.

Fig. 2. Shoots length ofL. perenne grown in raw(RM1, RM2), control (C) and mixed(RM1qC, RM2qC) soils.

different plant species and is attributed to a con-taminant fertilising effect(Von Arb and Brunold,1990; Carreras et al., 1998).The uptake of Cu, Pb and Zn in the shoots of

L. perenne (Table 6) was below the maximumtoxic levels to plants indicated by Kabata-Pendias

(2001) in mature leaf tissue, with metals taken upin the following order: Zn)Cu4Pb. The highestZn concentration was detected in the shoots ofplants grown on RM2qC soil (123 mg kg ) ay1

level that was significantly higher(40% and 55%)than that determined in RM2 and in C, respective-ly. A lower concentration of Zn, 92 mg kg , wasy1

found in shoots collected from RM1qC, withdifferences only slightly different respect to valuesdetected in the shoots of plants grown in RM1 andC substrates. Levels of Cu are comparable withthose measured in ryegrass grown on urban con-taminated soils by Sauve et al.(1996) and higher´in plants grown in RM1 and RM2(19.3 and 12.6mg kg , respectively), compared with RM1qCy1

(17.1 mg kg ), and RM2qC (10.4 mg kg ).y1 y1

The concentration of Pb was low, ranging from0.67 mg kg in RM2, to 1.03 mg kg in RM2qy1 y1

C.

3.4. Metal distribution in soil before and afterplanting

The sequential extraction scheme operationallysplit the total heavy metal distribution of the soilsinto a five-pool system. Although the specificityand selectivity of the extractants are often subject

21M. Arienzo et al. / The Science of the Total Environment 319 (2004) 13–25

Fig. 3. Chlorophylla(Cha), chlorophyll b(Chb) and total chlorophyll(Chtot) concentration in leaves ofL. perenne grown in raw(RM1, RM2), control (C) and mixed(RM1qC, RM2qC) soils. Values within the same column followed by the same letter arenot statistically significant atPs0.05.

Table 6Concentrations of metals in shoots ofL. perenne grown in raw(RM1, RM2), control (C) and mixed(RM1qC, RM2qC) soils.Values within the same column followed by the same letter are not statistically significant atPs0.05

Soil sample Cu Pb Znmg kg dry mattery1 mg kg dry mattery1 mg kg dry mattery1

RM1 19.3a 0.98b 99b(0.37) (0.27) (5.59)

RM2 12.6c 0.67b 88bc(0.89) (0.17) (8.24)

C 10.1d Bdl 79c(0.40) (0.95)

RM1qC (3:1 wyw) 17.1b 0.69b 92bc(0.45) (0.03) (3.11)

RM2qC (3:1 wyw) 10.4d 1.03b 123a(0.39) (0.4) (10)

Standard deviation in parentheses.

to criticism, the metal forms sequentially removedare: soluble and exchangeable(pool 1), carbonate-bound (pool 2), easily reducible metal oxide-bound (pool 3), organic-bound (pool 4) andresidual, mainly in well-crystallized oxides andsulfides and primary mineral lattice(pool 5). Theamounts of Cu, Pb and Zn in the sequentiallyextracted pools before and afterLolium growth arepresented in Fig. 4a and b, respectively. For allmetals a satisfactory agreement(r ranging between0.837 and 0.993;P-0.05) was found between thetotal obtained by a single determination(Table 4)and the cumulative totals as extracted by the

sequential extraction scheme before and afterLol-ium growth (Fig. 4a, b). In the majority of cases(90% of the values for Cu, 60% for Pb and 80%for Zn) mass balance of the sequential extractionswas negative. BeforeLolium, for Cu the cumula-tive totals measured fell withinq15yy22% ofthe gross total, in one sample(RM2) deviating byy37%. For Pb and Zn the cumulative total fellwithin q27yy25% andq5yy22% of the soilgross total, respectively. After plant growth, differ-ences between gross and cumulative totals werefor Cuy16yy48% with a higher value(y63%)in RM2, for Pbq32yy17% and for Znq29yy

22 M. Arienzo et al. / The Science of the Total Environment 319 (2004) 13–25

Fig. 4. Distribution of Cu, Pb and Zn in the raw(RM1, RM2), control (C), and mixed(RM1qC, RM2qC) soils before(a) andafter (b) L. perenne growth.

33%. Soil sample inhomogeneity, differences ofextraction efficiency and analytical errors can like-ly explain the deviations found between the totaland sequentially extracted metal contents(Adamoet al., 2002b).Chemical fractionation of metals beforeLolium

(Fig. 4a) indicated, for all five substrates, a mini-mal presence of Pb and Zn in the soluble andexchangeable pools. In the case of Cu, the contri-bution of this pool to the total ranged from 1.2%in RM1qC to 10.7% in RM2. Limited concentra-tions of metals in the more plant-accessible poolwere also found in the previous study carried outin the same area(Adamo et al., 2002a).

Copper in the C soil prevailed in the organicmatter pool(65%). In the RM1 and RM2 soilscopper was almost equally distributed betweencarbonate(21 and 34%), organic(35 and 40%),and residual(40 and 15%), forms. Mixing of thesesoils with C strongly affected the Cu organic pool,which increased to 59% in RM1qC, and to 64%in RM2qC. This was in agreement with thecommon view that in soil Cu is preferably asso-ciated with organic matter(McLaren and Craw-ford, 1973).Lead in both C and RM1 and RM2 soils was

mainly bound to organic matter(approx. 41%),followed by residual(approx. 27%), carbonate

23M. Arienzo et al. / The Science of the Total Environment 319 (2004) 13–25

(approx. 16%) and easily reducible oxides(approx. 16%), forms. No observable changes inmetal distribution were detected after mixing.Zinc in the C and RM2 soils was equally

distributed between carbonate(33 and 28%),organic(30 and 37%), and residual(36 and 34%),forms, while in the more contaminated RM1 soil71% of the metal occurred in the residual pool.Mixing of the soils did not significantly modifythese distributions.The addition of the C soil resulted in a distinct

redistribution among metal forms only for Cu.Speciation of trace elements in soil may vary,depending on the solid-phase components presentand their ability to release the metal, pH, numberand accessibility of adsorption sites and metalaffinity for soil components(Ritchie and Sposito,2002). In the studied soils, Cu is the metal moststrongly bound by organic matter. The increase oforganic carbon from 8 to 11 g kg with additiony1

of C soil to RM substrates increased organic Cuat the expense of exchangeable, carbonate andresidual forms. By contrast the C addition did notsignificantly affect Pb and Zn distribution presum-ably because they were more strongly bound byRM inorganic soil components and less easilycomplexed by organic matter.After plant growth(Fig. 4b) a general increase

in organic carbon content was observed, whichwas significant only in RM2(q34%), RM2qC(q35%) and RM1qC (q16%). Taking intoaccount the above-mentioned variability of data ofsequential and total extraction procedures as wellthe low metal uptake byLolium (Table 5), frac-tionation of Cu, Pb and Zn in soil after plantgrowth did not appear substantially different fromthat found before. Only a limited tendency todecrease the Cu and Zn organic pool, down to24%, was observed. This suggests in the shortperiod absence of plant-induced metal mobilityeffect according with the findings of otherresearchers(Ernst, 1996; Romkens and Dolfing,¨2002). However, as the extent of depletion ofmetal pools is related to the total plant uptake,according to Hammer and Keller(2002), thequestion remains whether a larger uptake(e.g. byrepeated planting or extended growth period)

might not result in a change of metal poolsdistribution.

3.5. Relationship between total metal content insoil, amounts removed by sequential extractionand plant uptake

Significant positive correlations were foundbetween soil(rs0.98; P-0.01) and plant(rs0.91; P-0.05) total metal content and the cumu-lative amount of Cu extracted beforeLolium bythe first four steps of sequential extraction(Pool1q2q3q4). A similar or greater correlation wasdetermined between total soil Zn content and thesum of the first four pools, when the values ofRM2 soil were excluded(rs0.99; P-0.001),whereas no relation was found with plant uptake.Specific chemical extractants used sequentiallytheoretically access different soil components,which bind and release heavy metals under differ-ent conditions and in response of plant uptake(Maand Uren, 1998). Taking into account that metaldistribution in soil strongly depends on soil prop-erties, metal involved and the type and age ofcontamination it might be likely that the appliedfour-step sequential extraction procedure quantifiesthe metal plant-available species in soil better forCu than for Zn and Pb.

4. Conclusions

The present study indicated thatL. perenne cangrow healthy on the soils of the IBSP site devotedto stacking raw materials, and especially well onthose derived from mixing of polluted with non-polluted soil. Data also indicate that when metalavailability to the plant was low, even a non-tolerant plant species, likeL. perenne, was able tosurvive. Therefore, the above results support fur-ther investigations on the possibility of applyingthe proposed rehabilitation strategy on a largerscale. This could represent an economically morerealistic and cost-effective alternative than exca-vation and off site disposal or revegetation incombination with metal-immobilizing additives,especially for a vast industrial site such as IBSP.According to our preliminary results metals seemto remain stable over the time period studied(3

24 M. Arienzo et al. / The Science of the Total Environment 319 (2004) 13–25

months) and this will be followed during thecoming years. It can be anticipated that furtherdevelopment of the vegetation will produce ahumus layer, which is currently absent. Nascenthumus is expected to result in a supplementarymetal-immobilizing capacity of the upper soillayer.

Acknowledgments

The authors wish to thank Mr Sergio Carotenutofor his technical assistance in greenhouse trials.The authors are also grateful to anonymous refe-rees and to the STOTEN Associate Editor Dr J.P.Bennett for their comments, which improved thepresentation of the results, and for the helpfulreview of the manuscript. This study was finan-cially supported by Campania Region(POP funds,action 5.4.2, 1999). This paper represents JournalSeries N. 0043 from the DISSPA.

References

Adamo P, Arienzo M, Bianco MR, Violante P. Heavy metalcontamination of the soils used for stocking raw materialsin the former ILVA iron-steel industrial plant of Bagnoli(Italy). Sci Total Environ 2002a;295:17–34.

Adamo P, Dudka S, Wilson MJ, McHardy WJ. Distribution oftrace elements in soils from the Sudbury smelting area(Ontario, Canada). Water Air Soil Pollut 2002b;135:1–22.

Buondonno C, Ermice A, Buondonno A, Murolo M, PuglianoML. Human-influenced soils from an iron and steel worksin Naples, Italy. Soil Sci Soc Am J 1998;62:694–700.

Carreras HA, Gudino GL, Pignata ML. Comparative biomon-˜itoring of atmospheric quality in five zones of Cordoba city(Argentina) employing the transplanted lichenUsnea sp.Environ Pollut 1998;103:317–325.

De Boodt MF. Application of the sorption theory to eliminateheavy metals from waste waters and contaminated soils. In:Bolt GH, De Boodt MF, Hayes MBH, McBride MB, editors.NATO ASI series E: applied sciences, vol. 190. Dordrecht:Kluwer Academic Publishers, 1991. p. 293–320.

Ernst WHO. Bioavailability of heavy metals and decontami-nation of soils by plants. Appl Geochem 1996;11:163–167.

Gworek B. Inactivation of cadmium in contaminated soilsusing synthetic zeolites. Environ Pollut 1992;75:269–271.

Hammer D, Keller C. Changes in the rhizosphere of metal-accumulating plants evidenced by chemical extractants. JEnviron Qual 2002;31:1561–1569.

Hendershot WH, Duquette M. A simple barium chloridemethod for determining cation exchange capacity andexchangeable cations. Soil Sci Soc Am J 1986;50:606–608.

Italian Ministry of Environment. DM471-Regolamento recantecriteri, procedure e modalita per la messa in sicurezza, la`bonifica e il ripristino ambientale dei siti inquinati. Supple-mento Ordinario Gazzetta Ufficiale della Repubblica ItalianaN. 293 del 15y12y1999: 67.

Kabata-Pendias A. Trace elements in soil and plant. Boca-Raton, Florida: CRC Press, 2001. p. 413

Lichtenthaler HK. Chlorophylls and carotenoids: pigments ofphotosynthetic biomembranes. Methods Enzymol1987;148:350–382.

Loeppert RH, Hallmark CT, Koshy MM. Routine procedurefor rapid determination of soil carbonates. Soil Sci Soc AmJ 1984;48:1030–1033.

Ma Y, Uren N. Transformation of heavy metals added to soil-Application of new sequential extraction procedure. Geo-derma 1998;84:157–168.

Marschner H. Mineral Nutrition of higher plants. London:Academic Press, 1986.

McGrath SP, Shen ZG, Zhao FJ. Heavy metal uptake andchemical changes in the rhizosphere of Thlaspi caerulescensand Thlaspi ochroleucum grown in contaminated soils. PlantSoil 1997;188:153–159.

McLaren RG, Crawford DV. Studies on soil copper. I. Thefractionation of copper in soils. J Soil Sci 1973;24:172–181.

Nelson DW, Sommers LE. Total carbon, organic carbon andorganic matter. In: Page AL, et al, editor. Methods of soilanalysis. Part 2, 2nd ed. Madison, WI: ASA and SSA, 1982.p. 439–580 (Agron. Monogr. 9).

Norland MR, Veith DL. Revegetation of coarse taconite ironore tailing using municipal solid waste compost. J HazardMater 1995;41:123–134.

Olsen SR, Sommers LE. Phosphorus. In: Page AL, et al,editor. Methods of soil analysis. Part 2, 2nd ed. Madison,WI: ASA and SSA, 1982. p. 403–430 (Agron. Monogr. 9).

Pichtel J, Salt CA. Vegetative growth and trace metal accu-mulation on metalliferous waste. J Environ Qual1998;27:618–624.

Ritchie GSP, Sposito G. Speciation in soils. In: Ure AM,Davidson CM, editors. Chemical speciation in the environ-ment. Oxon: Blackwell Science, 2002. p. 237–264.

Romkens PFAM, Dolfing J. Effect of Ca on the solubility and¨molecular size distribution of DOC and Cu binding in soilsolution samples. Environ Sci Technol 2002;32:363–369.

Sauve S, Cook N, Hendershot WH, McBride MB. Linking´plant tissue concentrations and soil copper pools in urbancontaminated soils. Environ Pollut 1996;94:153–157.

Smith RAH, Bradshaw AD. The use of metal tolerant plantpopulations for the reclamation of metalliferous wastes.Appl Ecol 1979;16:595–612.

Testa SM. The reuse and recycling of contaminated soil. Bocaraton: CRC, 1997. p. 268

USEPA. Federal Register, Thursday, March 29, 1990, 40 CFRPart 261.

Vangronsveld J, Clijsters H. Toxic effects of metals. In: FaragoME, editor. Plants and the chemical elements. Biochemistry,

25M. Arienzo et al. / The Science of the Total Environment 319 (2004) 13–25

uptake, tolerance and toxicity. Weinheim: VCH Verlagsge-sellshaft, 1994. p. 149–177.

Vangronsveld J, Van Assche F, Clijsters H. Reclamation of a‘desert like’ site in the North east of Belgium: Evolution ofthe metal pollution and experiments in situ. In: Farmer JG,editor. Proceedings of the International Conference of HeavyMetals in the Environment, Edinburgh: CEP Consultants,1991:58–61.

Vangronsveld J, Van Assche F, Sterckx J, Clijsters H. Reha-bilitation studies on an old non-ferrous waste dumpingground: effects of metal immobilization and revegetation.In: Allen RJ, Nriagu JO, editors. Proceedings of the Inter-national Conference of Heavy Metals in the Environment,Edinburgh: CEP Consultants, 1993:563–566.

Vangronsveld J, Van Assche F, Clijsters H. Reclamation of abare industrial area contaminated by non-ferrous metals: in

situ metal immobilization and revegetation. Environ Pollut1995;87:51–59.

Vangronsveld J, Mocquot B, Mench M, Clijsters H. Utilisationspratiques de biomarqueurs pour l’etude de la contamination´des ecosystems par les principales classes de micropolluants.Biomarqueurs d’exposition des vegetaux terrestres aux pol-´ ´luants. Metaux. In: Lagadic L, Amiard JC, Caquet T,´Ramade F, editors. Utilisation de biomarquers en ecotoxi-cologie. Paris, France: Masson, 1996.

Von Arb C, Brunold C. Lichen physiology and air pollution.I. Physiological responses of in situParmelia sulcata amongair pollution zones within Biel, Switzerland. Can J Bot1990;68:35–42.

Zar JH. Biostatistical analysis. Englewood Cliffs, NJ: PrenticeHall, 1984.